Acute ischemic stroke in tuberculous meningitis.

Background: The underlying mechanism for stroke in patients with tuberculous meningitis (TBM) remains unclear. This study aimed to investigate the predictors of acute ischemic stroke (AIS) in TBM and whether AIS mediates the relationship between inflammation markers and functional disability. Methods: TBM patients admitted to five hospitals between January 2011 and December 2021 were consecutively observed. Generalized linear mixed model and subgroup analyses were performed to investigate predictors of AIS in patients with and without vascular risk factors (VAFs). Mediation analyses were performed to explore the potential causal chain in which AIS may mediate the relationship between neuroimaging markers of inflammation and 90-day functional outcomes. Results: A total of 1,353 patients with TBM were included. The percentage rate of AIS within 30 days after admission was 20.4 (95% CI, 18.2-22.6). A multivariate analysis suggested that age ≥35 years (OR = 1.49; 95% CI, 1.06-2.09; P = 0.019), hypertension (OR = 3.56; 95% CI, 2.42-5.24; P < 0.001), diabetes (OR = 1.78; 95% CI, 1.11-2.86; P = 0.016), smoking (OR = 2.88; 95% CI, 1.68-4.95; P < 0.001), definite TBM (OR = 0.19; 95% CI, 0.06-0.42; P < 0.001), disease severity (OR = 2.11; 95% CI, 1.50-2.90; P = 0.056), meningeal enhancement (OR = 1.66; 95% CI, 1.19-2.31; P = 0.002), and hydrocephalus (OR = 2.98; 95% CI, 1.98-4.49; P < 0.001) were associated with AIS. Subgroup analyses indicated that disease severity (P for interaction = 0.003), tuberculoma (P for interaction = 0.008), and meningeal enhancement (P for interaction < 0.001) were significantly different in patients with and without VAFs. Mediation analyses revealed that the proportion of the association between neuroimaging markers of inflammation and functional disability mediated by AIS was 16.98% (95% CI, 7.82-35.12) for meningeal enhancement and 3.39% (95% CI, 1.22-6.91) for hydrocephalus. Conclusion: Neuroimaging markers of inflammation were predictors of AIS in TBM patients. AIS mediates < 20% of the association between inflammation and the functional outcome at 90 days. More attention should be paid to clinical therapies targeting inflammation and hydrocephalus to directly improve functional outcomes.

Multi-omics analysis reveals that linoleic acid metabolism is associated with variations of trained immunity induced by distinct BCG strains.

Trained immunity is one of the mechanisms by which BCG vaccination confers persistent nonspecific protection against diverse diseases. Genomic differences between the different BCG vaccine strains that are in global use could result in variable protection against tuberculosis and therapeutic effects on bladder cancer. In this study, we found that four representative BCG strains (BCG-Russia, BCG-Sweden, BCG-China, and BCG-Pasteur) covering all four genetic clusters differed in their ability to induce trained immunity and nonspecific protection. The trained immunity induced by BCG was associated with the Akt-mTOR-HIF1α axis, glycolysis, and NOD-like receptor signaling pathway. Multi-omics analysis (epigenomics, transcriptomics, and metabolomics) showed that linoleic acid metabolism was correlated with the trained immunity-inducing capacity of different BCG strains. Linoleic acid participated in the induction of trained immunity and could act as adjuvants to enhance BCG-induced trained immunity, revealing a trained immunity-inducing signaling pathway that could be used in the adjuvant development.

A nursing report on a corneal contact lens wearer receiving keratoplasty due to corneal ulcer and perforation caused by Pythium insidiosum infection: A case report.

BACKGROUND: To report the nursing experience of a case of corneal contact lens wearer receiving the 2nd keratoplasty due to corneal ulcer and perforation caused by Pythium insidiosum infection. METHODS: A 30-year-old female patient had blurred vision after deep anterior lamellar keratoplasty for a right corneal ulcer. At the 5th week, the right eye appeared the symptoms, such as redness and pain. The anterior segment photography was performed on the eye, and the result showed that the epithelium was missing in the right eye lesion area, and a large number of longitudinal and transversal streaks were visible from the epithelium to the stroma, with fungus filaments to be discharged. Upon macro-genome sequencing of the corneal secretion, a P. insidiosum infection was observed. Then, the patient underwent the keratoplasty, and 3 weeks later, the corneal implant showed a tendency to dissolve, the sutures were partially loosened, and the eye was almost blind. Subsequently, the patient was admitted to our hospital and subject to the 2nd penetrating keratoplasty of the right eye (allograft). After surgery, linezolid and azithromycin injections were given through intravenous drip and local drip of the eye for anti-inflammation, and tacrolimus eye drops for antirejection. RESULTS: Postoperatively, the patient showed signs of recovery with slight corneal edema and visible pupil, leading to discharge with improved vision. The corneal implant was normal 1 week after surgery and the vision of the right eye was hand move/before eye at the 6th month of follow-up. Continuous care and removal of sutures 3 months post-surgery contributed to a successful outcome, with the patient achieving hand motion vision 6 months after the procedure. CONCLUSION: Corneal ulcer caused by P. insidiosum infection not only needs timely and effective keratoplasty intervention, but also requires perfect nursing measures.

Citriquinolinones A and B: Rare Isoquinolinone-Embedded Citrinin Analogues and Related Metabolites from the Deep-Sea-Derived Aspergillus versicolor 170217.

A systematic chemical investigation of the deep-sea-derived fungus Aspergillus versicolor 170217 resulted in the isolation of six new (1-6) and 45 known (7-51) compounds. The structures of the new compounds were established on the basis of exhaustive analysis of their spectroscopic data and theoretical-statistical approaches including GIAO-NMR, TDDFT-ECD/ORD calculations, DP4+ probability analysis, and biogenetic consideration. Citriquinolinones A (1) and B (2) feature a unique isoquinolinone-embedded citrinin scaffold, representing the first exemplars of a citrinin-isoquinolinone hybrid. Dicitrinones K-L (3-4) are two new dimeric citrinin analogues with a rare CH-CH3 bridge. Biologically, frangula-emodin (32) and diorcinol (17) displayed remarkable anti-food allergic activity with IC50 values of 7.9 ± 3.0 µM and 13.4 ± 1.2 µM, respectively, while diorcinol (17) and penicitrinol A (20) exhibited weak inhibitory activity against Vibrio parahemolyticus, with MIC values ranging from 128 to 256 µM.

Efficacy and Safety of Ferric Citrate on Hyperphosphatemia among Chinese Patients with Chronic Kidney Disease Undergoing Hemodialysis: A Phase III Multicenter Randomized Open-Label Active-Drug-Controlled Study.

INTRODUCTION: Hyperphosphatemia in chronic kidney disease (CKD) patients is positively associated with mortality. Ferric citrate is a potent phosphorus binder that lowers serum phosphorus level and improves iron metabolism. We compared its efficacy and safety with active drugs in Chinese CKD patients with hemodialysis. METHODS: Chinese patients undergoing hemodialysis were randomized into two treatment groups in a 1:1 ratio, receiving either ferric citrate or sevelamer carbonate, respectively, for 12 weeks. Serum phosphorus levels, calcium concentration, and iron metabolism parameters were evaluated every 2 weeks. Frequency and severity of adverse events were recorded. RESULTS: 217 (90.4%) patients completed the study with balanced demographic and baseline characteristics between two groups. Ferric citrate decreased the serum phosphorus level to 0.59 ± 0.54 mmol/L, comparable to 0.56 ± 0.62 mmol/L by sevelamer carbonate. There was no significant difference between two groups (p > 0.05) in the proportion of patients with serum phosphorus levels reaching the target range, the response rate to the study drug, and the changes of corrected serum calcium concentrations, and intact-PTH levels at the end of treatment. The change of iron metabolism indicators in the ferric citrate group was significantly higher than those in the sevelamer carbonate group. There are 47 (40.5%) patients in the ferric citrate group, and 26 (21.3%) patients in the sevelamer carbonate group experienced drug-related treatment emergent adverse events (TEAEs); most were mild and tolerable. Common drug-related TEAEs were gastrointestinal disorders, including diarrhea (12.9 vs. 2.5%), fecal discoloration (14.7 vs. 0%), and constipation (1.7 vs. 7.4%) in ferric citrate and sevelamer carbonate group. CONCLUSION: Ferric citrate capsules have good efficacy and safety in the control of hyperphosphatemia in adult patients with CKD undergoing hemodialysis. Efficacy is not inferior to sevelamer carbonate. The TEAEs were mostly mild and tolerated by the patients.

Prognostic value of blood urea nitrogen to serum albumin ratio for acute kidney injury and in-hospital mortality in intensive care unit patients with intracerebral haemorrhage: a retrospective cohort study using the MIMIC-IV database.

OBJECTIVE: We sought to evaluate the prognostic ability of blood urea nitrogen to serum albumin ratio (BAR) for acute kidney injury (AKI) and in-hospital mortality in patients with intracerebral haemorrhage (ICH) in intensive care unit (ICU). DESIGN: A retrospective cohort study using propensity score matching. SETTING: ICU of Beth Israel Deaconess Medical Center. PARTICIPANTS: The data of patients with ICH were obtained from the Medical Information Mart for Intensive Care IV (V.1.0) database. A total of 1510 patients with ICH were enrolled in our study. MAIN OUTCOME AND MEASURE: The optimal threshold value of BAR is determined by the means of X-tile software (V.3.6.1) and the crude cohort was categorised into two groups on the foundation of the optimal cut-off BAR (6.0 mg/g). Propensity score matching and inverse probability of treatment weighting were performed to control for confounders. The predictive performance of BAR for AKI was tested using univariate and multivariate logistic regression analyses. Multivariate Cox regression analysis was used to investigate the association between BAR and in-hospital mortality. RESULTS: The optimal cut-off value for BAR was 6.0 mg/g. After matching, multivariate logistic analysis showed that the high-BAR group had a significantly higher risk of AKI (OR, 2.60; 95% confidence index, 95% CI, 1.86 to 3.65, p<0.001). What's more, a higher BAR was also an independent risk factor for in-hospital mortality (HR, 2.84; 95% confidence index, 95% CI, 1.96 to 4.14, p<0.001) in terms of multivariate Cox regression analysis. These findings were further demonstrated in the validation cohort. CONCLUSIONS: BAR is a promising and easily available biomarker that could serve as a prognostic predictor of AKI and in-hospital mortality in patients with ICH in the ICU.

A Survey of Convolutional Neural Network in Breast Cancer.

Problems: For people all over the world, cancer is one of the most feared diseases. Cancer is one of the major obstacles to improving life expectancy in countries around the world and one of the biggest causes of death before the age of 70 in 112 countries. Among all kinds of cancers, breast cancer is the most common cancer for women. The data showed that female breast cancer had become one of the most common cancers. Aims: A large number of clinical trials have proved that if breast cancer is diagnosed at an early stage, it could give patients more treatment options and improve the treatment effect and survival ability. Based on this situation, there are many diagnostic methods for breast cancer, such as computer-aided diagnosis (CAD). Methods: We complete a comprehensive review of the diagnosis of breast cancer based on the convolutional neural network (CNN) after reviewing a sea of recent papers. Firstly, we introduce several different imaging modalities. The structure of CNN is given in the second part. After that, we introduce some public breast cancer data sets. Then, we divide the diagnosis of breast cancer into three different tasks: 1. classification; 2. detection; 3. segmentation. Conclusion: Although this diagnosis with CNN has achieved great success, there are still some limitations. (i) There are too few good data sets. A good public breast cancer dataset needs to involve many aspects, such as professional medical knowledge, privacy issues, financial issues, dataset size, and so on. (ii) When the data set is too large, the CNN-based model needs a sea of computation and time to complete the diagnosis. (iii) It is easy to cause overfitting when using small data sets.

GAM-SpCaNet: Gradient awareness minimization-based spinal convolution attention network for brain tumor classification.

Brain tumor is one of the common diseases of the central nervous system, with high morbidity and mortality. Due to the wide range of brain tumor types and pathological types, the same type is divided into different subgrades. The imaging manifestations are complex, making clinical diagnosis and treatment difficult. In this paper, we construct SpCaNet (Spinal Convolution Attention Network) to effectively utilize the pathological features of brain tumors, consisting of a Positional Attention (PA) convolution block, Relative self-attention transformer block, and Intermittent fully connected (IFC) layer. Our method is more lightweight and efficient in recognition of brain tumors. Compared with the SOTA model, the number of parameters is reduced by more than three times. In addition, we propose the gradient awareness minimization (GAM) algorithm to solve the problem of insufficient generalization ability of the traditional Stochastic Gradient Descent (SGD) method and use it to train the SpCaNet model. Compared with SGD, GAM achieves better classification performance. According to the experimental results, our method has achieved the highest accuracy of 99.28%, and the proposed method performs well in classifying brain tumors.

Automated diagnosis of cardiovascular diseases from cardiac magnetic resonance imaging using deep learning models: A review.

In recent years, cardiovascular diseases (CVDs) have become one of the leading causes of mortality globally. At early stages, CVDs appear with minor symptoms and progressively get worse. The majority of people experience symptoms such as exhaustion, shortness of breath, ankle swelling, fluid retention, and other symptoms when starting CVD. Coronary artery disease (CAD), arrhythmia, cardiomyopathy, congenital heart defect (CHD), mitral regurgitation, and angina are the most common CVDs. Clinical methods such as blood tests, electrocardiography (ECG) signals, and medical imaging are the most effective methods used for the detection of CVDs. Among the diagnostic methods, cardiac magnetic resonance imaging (CMRI) is increasingly used to diagnose, monitor the disease, plan treatment and predict CVDs. Coupled with all the advantages of CMR data, CVDs diagnosis is challenging for physicians as each scan has many slices of data, and the contrast of it might be low. To address these issues, deep learning (DL) techniques have been employed in the diagnosis of CVDs using CMR data, and much research is currently being conducted in this field. This review provides an overview of the studies performed in CVDs detection using CMR images and DL techniques. The introduction section examined CVDs types, diagnostic methods, and the most important medical imaging techniques. The following presents research to detect CVDs using CMR images and the most significant DL methods. Another section discussed the challenges in diagnosing CVDs from CMRI data. Next, the discussion section discusses the results of this review, and future work in CVDs diagnosis from CMR images and DL techniques are outlined. Finally, the most important findings of this study are presented in the conclusion section.

Diagnosis of cerebral microbleed via VGG and extreme learning machine trained by Gaussian map bat algorithm.

Cerebral microbleed (CMB) is a serious public health concern. It is associated with dementia, which can be detected with brain magnetic resonance image (MRI). CMBs often appear as tiny round dots on MRIs, and they can be spotted anywhere over brain. Therefore, manual inspection is tedious and lengthy, and the results are often short in reproducible. In this paper, a novel automatic CMB diagnosis method was proposed based on deep learning and optimization algorithms, which used the brain MRI as the input and output the diagnosis results as CMB and non-CMB. Firstly, sliding window processing was employed to generate the dataset from brain MRIs. Then, a pre-trained VGG was employed to obtain the image features from the dataset. Finally, an ELM was trained by Gaussian-map bat algorithm (GBA) for identification. Results showed that the proposed method VGG-ELM-GBA provided better generalization performance than several state-of-the-art approaches.

Multiple-level thresholding for breast mass detection.

Detection of breast mass plays a very important role in making the diagnosis of breast cancer. For faster detection of breast cancer caused by breast mass, we developed a novel and efficient patch-based breast mass detection system for mammography images. The proposed framework is comprised of three modules, including pre-processing, multiple-level breast tissue segmentation, and final breast mass detection. An improved Deeplabv3+ model for pectoral muscle removal is deployed in pre-processing. We then proposed a multiple-level thresholding segmentation method to segment breast mass and obtained the connected components (ConCs), where the corresponding image patch to each ConC is extracted for mass detection. In the final detection stage, each image patch is classified into breast mass and breast tissue background by trained deep learning models. The patches that are classified as breast mass are then taken as the candidates for breast mass. To reduce the false positive rate in the detection results, we applied the non-maximum suppression algorithm to combine the overlapped detection results. Once an image patch is considered a breast mass, the accurate detection result can then be retrieved from the corresponding ConC in the segmented images. Moreover, a coarse segmentation result can be simultaneously retrieved after detection. Compared to the state-of-the-art methods, the proposed method achieved comparable performance. On CBIS-DDSM, the proposed method achieved a detection sensitivity of 0.87 at 2.86 FPI (False Positive rate per Image), while the sensitivity reached 0.96 on INbreast with an FPI of only 1.29.

CTMLP: Can MLPs replace CNNs or transformers for COVID-19 diagnosis?

BACKGROUND: Convolutional Neural Networks (CNNs) and the hybrid models of CNNs and Vision Transformers (VITs) are the recent mainstream methods for COVID-19 medical image diagnosis. However, pure CNNs lack global modeling ability, and the hybrid models of CNNs and VITs have problems such as large parameters and computational complexity. These models are difficult to be used effectively for medical diagnosis in just-in-time applications. METHODS: Therefore, a lightweight medical diagnosis network CTMLP based on convolutions and multi-layer perceptrons (MLPs) is proposed for the diagnosis of COVID-19. The previous self-supervised algorithms are based on CNNs and VITs, and the effectiveness of such algorithms for MLPs is not yet known. At the same time, due to the lack of ImageNet-scale datasets in the medical image domain for model pre-training. So, a pre-training scheme TL-DeCo based on transfer learning and self-supervised learning was constructed. In addition, TL-DeCo is too tedious and resource-consuming to build a new model each time. Therefore, a guided self-supervised pre-training scheme was constructed for the new lightweight model pre-training. RESULTS: The proposed CTMLP achieves an accuracy of 97.51%, an f1-score of 97.43%, and a recall of 98.91% without pre-training, even with only 48% of the number of ResNet50 parameters. Furthermore, the proposed guided self-supervised learning scheme can improve the baseline of simple self-supervised learning by 1%-1.27%. CONCLUSION: The final results show that the proposed CTMLP can replace CNNs or Transformers for a more efficient diagnosis of COVID-19. In addition, the additional pre-training framework was developed to make it more promising in clinical practice.

ReRNet: A Deep Learning Network for Classifying Blood Cells.

AIMS: Blood cell classification helps detect various diseases. However, the current classification model of blood cells cannot always get great results. A network that automatically classifies blood cells can provide doctors with data as one of the criteria for diagnosing patients' disease types and severity. If doctors diagnose blood cells, doctors could spend lots of time on the diagnosis. The diagnosis progress is very tedious. Doctors can make some mistakes when they feel tired. On the other hand, different doctors may have different points on the same patient. METHODS: We propose a ResNet50-based ensemble of randomized neural networks (ReRNet) for blood cell classification. ResNet50 is used as the backbone model for feature extraction. The extracted features are fed to 3 randomized neural networks (RNNs): Schmidt neural network, extreme learning machine, and dRVFL. The outputs of the ReRNet are the ensemble of these 3 RNNs based on the majority voting mechanism. The 5 × 5-fold cross-validation is applied to validate the proposed network. RESULTS: The average-accuracy, average-sensitivity, average-precision, and average-F1-score are 99.97%, 99.96%, 99.98%, and 99.97%, respectively. CONCLUSIONS: The ReRNet is compared with 4 state-of-the-art methods and achieves the best classification performance. The ReRNet is an effective method for blood cell classification based on these results.

An integrated approach of network pharmacology, molecular docking, and experimental verification uncovers kaempferol as the effective modulator of HSD17B1 for treatment of endometrial cancer.

BACKGROUND: Endometrial cancer (EC) is one of the most common gynecological malignancies globally, and the development of innovative, effective drugs against EC remains a key issue. Phytoestrogen kaempferol exhibits anti-cancer effects, but the action mechanisms are still unclear. METHOD: MTT assays, colony-forming assays, flow cytometry, scratch healing, and transwell assays were used to evaluate the proliferation, apoptosis, cell cycle, migration, and invasion of both ER-subtype EC cells. Xenograft experiments were used to assess the effects of kaempferol inhibition on tumor growth. Next-generation RNA sequencing was used to compare the gene expression levels in vehicle-treated versus kaempferol-treated Ishikawa and HEC-1-A cells. A network pharmacology and molecular docking technique were applied to identify the anti-cancer mechanism of kaempferol, including the building of target-pathway network. GO analysis and KEGG pathway enrichment analysis were used to identify cancer-related targets. Finally, the study validated the mRNA and protein expression using real-time quantitative PCR, western blotting, and immunohistochemical analysis. RESULTS: Kaempferol was found to suppress the proliferation, promote apoptosis, and limit the tumor-forming, scratch healing, invasion, and migration capacities of EC cells. Kaempferol inhibited tumor growth and promotes apoptosis in a human endometrial cancer xenograft mouse model. No significant toxicity of kaempferol was found in human monocytes and normal cell lines at non-cytotoxic concentrations. No adverse effects or significant changes in body weight or organ coefficients were observed in 3-7 weeks' kaempferol-treated animals. The RNA sequencing, network pharmacology, and molecular docking approaches identified the overall survival-related differentially expressed gene HSD17B1. Interestingly, kaempferol upregulated HSD17B1 expression and sensitivity in ER-negative EC cells. Kaempferol differentially regulated PPARG expression in EC cells of different ER subtypes, independent of its effect on ESR1. HSD17B1 and HSD17B1-associated genes, such as ESR1, ESRRA, PPARG, AKT1, and AKR1C1\2\3, were involved in several estrogen metabolism pathways, such as steroid binding, 17-beta-hydroxysteroid dehydrogenase (NADP+) activity, steroid hormone biosynthesis, and regulation of hormone levels. The molecular basis of the effects of kaempferol treatment was evaluated. CONCLUSIONS: Kaempferol is a novel therapeutic candidate for EC via HSD17B1-related estrogen metabolism pathways. These results provide new insights into the efficiency of the medical translation of phytoestrogens.

Aculeaxanthones A-E, new xanthones from the marine-derived fungus Aspergillus aculeatinus WHUF0198.

Introduction: Dimeric natural products are widespread in plants and microorganisms, which usually have complex structures and exhibit greater bioactivities than their corresponding monomers. In this study, we report five new dimeric tetrahydroxanthones, aculeaxanthones A-E (4-8), along with the homodimeric tetrahydroxanthone secalonic acid D (1), chrysoxanthones B and C (2 and 3), and 4-4'-secalonic acid D (9), from different fermentation batches of the title fungus. Methods: A part of the culture was added to a total of 60 flasks containing 300 ml each of number II fungus liquid medium and culture 4 weeks in a static state at 28˚C. The liquid phase (18 L) and mycelia was separated from the fungal culture by filtering. A crude extract was obtained from the mycelia by ultrasound using acetone. To obtain a dry extract (18 g), the liquid phase combined with the crude extract were further extracted by EtOAc and concentrated in vacuo. The MIC of anaerobic bacteria was examined by a broth microdilution assay. To obtain MICs for aerobic bacteria, the agar dilution streak method recommended in Clinical and Laboratory Standards Institute document (CLSI) M07-A10 was used. Compounds 1-9 was tested against the Bel-7402, A-549 and HCT-116 cell lines according to MTT assay. Results and Discussion: The structures of these compounds were elucidated on the base of 1D and 2D NMR and HR-ESIMS data, and the absolute configurations of the new xanthones 4-8 were determined by conformational analysis and time-dependent density functional theory-electronic circular dichroism (TDDFT-ECD) calculations. Compounds 1-9 were tested for cytotoxicity against the Bel-7402, A549, and HCT-116 cancer cell lines. Of the dimeric tetrahydroxanthone derivatives, only compound 6 provided cytotoxicity effect against Bel-7402 cell line (IC50, 1.96 µM). Additionally, antimicrobial activity was evaluated for all dimeric tetrahydroxanthones, including four Gram-positive bacteria including Enterococcus faecium ATCC 19434, Bacillus subtilis 168, Staphylococcus aureus ATCC 25923 and MRSA USA300; four Gram-negative bacteria, including Helicobacter pylori 129, G27, as well as 26,695, and multi drug-resistant strain H. pylori 159, and one Mycobacterium M. smegmatis ATCC 607. However, only compound 1 performed activities against H. pylori G27, H. pylori 26695, H. pylori 129, H. pylori 159, S. aureus USA300, and B. subtilis 168 with MIC values of 4.0, 4.0, 2.0, 2.0, 2.0 and 1.0 µg/mL, respectively.

An Evolutionary Attention-Based Network for Medical Image Classification.

Deep learning has become a primary choice in medical image analysis due to its powerful representation capability. However, most existing deep learning models designed for medical image classification can only perform well on a specific disease. The performance drops dramatically when it comes to other diseases. Generalizability remains a challenging problem. In this paper, we propose an evolutionary attention-based network (EDCA-Net), which is an effective and robust network for medical image classification tasks. To extract task-related features from a given medical dataset, we first propose the densely connected attentional network (DCA-Net) where feature maps are automatically channel-wise weighted, and the dense connectivity pattern is introduced to improve the efficiency of information flow. To improve the model capability and generalizability, we introduce two types of evolution: intra- and inter-evolution. The intra-evolution optimizes the weights of DCA-Net, while the inter-evolution allows two instances of DCA-Net to exchange training experience during training. The evolutionary DCA-Net is referred to as EDCA-Net. The EDCA-Net is evaluated on four publicly accessible medical datasets of different diseases. Experiments showed that the EDCA-Net outperforms the state-of-the-art methods on three datasets and achieves comparable performance on the last dataset, demonstrating good generalizability for medical image classification.

WACPN: A Neural Network for Pneumonia Diagnosis.

Community-acquired pneumonia (CAP) is considered a sort of pneumonia developed outside hospitals and clinics. To diagnose community-acquired pneumonia (CAP) more efficiently, we proposed a novel neural network model. We introduce the 2-dimensional wavelet entropy (2d-WE) layer and an adaptive chaotic particle swarm optimization (ACP) algorithm to train the feed-forward neural network. The ACP uses adaptive inertia weight factor (AIWF) and Rossler attractor (RA) to improve the performance of standard particle swarm optimization. The final combined model is named WE-layer ACP-based network (WACPN), which attains a sensitivity of 91.87±1.37%, a specificity of 90.70±1.19%, a precision of 91.01±1.12%, an accuracy of 91.29±1.09%, F1 score of 91.43±1.09%, an MCC of 82.59±2.19%, and an FMI of 91.44±1.09%. The AUC of this WACPN model is 0.9577. We find that the maximum deposition level chosen as four can obtain the best result. Experiments demonstrate the effectiveness of both AIWF and RA. Finally, this proposed WACPN is efficient in diagnosing CAP and superior to six state-of-the-art models. Our model will be distributed to the cloud computing environment.

ELUCNN for explainable COVID-19 diagnosis.

COVID-19 is a positive-sense single-stranded RNA virus caused by a strain of coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Several noteworthy variants of SARS-CoV-2 were declared by WHO as Alpha, Beta, Gamma, Delta, and Omicron. Till 13/Dec/2022, it has caused 6.65 million death tolls, and over 649 million confirmed positive cases. Based on the convolutional neural network (CNN), this study first proposes a ten-layer CNN as the backbone model. Then, the exponential linear unit (ELU) is introduced to replace ReLU, and the traditional convolutional block is now transformed into conv-ELU. Finally, an ELU-based CNN (ELUCNN) model is proposed for COVID-19 diagnosis. Besides, the MDA strategy is used to enhance the size of the training set. We develop a mobile app integrating ELUCNN, and this web app is run on a client-server modeled structure. Ten runs of the tenfold cross-validation experiment show our model yields a sensitivity of 94.41 ± 0.98 , a specificity of 94.84 ± 1.21 , an accuracy of 94.62 ± 0.96 , and an F1 score of 94.61 ± 0.95 . The ELUCNN model and mobile app are effective in COVID-19 diagnosis and give better results than 14 state-of-the-art COVID-19 diagnosis models concerning accuracy.

PSTCNN: Explainable COVID-19 diagnosis using PSO-guided self-tuning CNN.

Since 2019, the coronavirus disease-19 (COVID-19) has been spreading rapidly worldwide, posing an unignorable threat to the global economy and human health. It is a disease caused by severe acute respiratory syndrome coronavirus 2, a single-stranded RNA virus of the genus Betacoronavirus. This virus is highly infectious and relies on its angiotensin-converting enzyme 2-receptor to enter cells. With the increase in the number of confirmed COVID-19 diagnoses, the difficulty of diagnosis due to the lack of global healthcare resources becomes increasingly apparent. Deep learning-based computer-aided diagnosis models with high generalisability can effectively alleviate this pressure. Hyperparameter tuning is essential in training such models and significantly impacts their final performance and training speed. However, traditional hyperparameter tuning methods are usually time-consuming and unstable. To solve this issue, we introduce Particle Swarm Optimisation to build a PSO-guided Self-Tuning Convolution Neural Network (PSTCNN), allowing the model to tune hyperparameters automatically. Therefore, the proposed approach can reduce human involvement. Also, the optimisation algorithm can select the combination of hyperparameters in a targeted manner, thus stably achieving a solution closer to the global optimum. Experimentally, the PSTCNN can obtain quite excellent results, with a sensitivity of 93.65%±1.86%, a specificity of 94.32%±2.07%, a precision of 94.30%±2.04%, an accuracy of 93.99%±1.78%, an F1-score of 93.97%±1.78%, Matthews Correlation Coefficient of 87.99%±3.56%, and Fowlkes-Mallows Index of 93.97%±1.78%. Our experiments demonstrate that compared to traditional methods, hyperparameter tuning of the model using an optimisation algorithm is faster and more effective.